PROCEEDINGS OF THE ARISTOTELIAN SOCIETY HOSTING AND PUBLISHING TALKS IN PHILOSOPHY SINCE 1880 The Linguistic Approach to Ontology LEE WALTERS 2020–2021 141ST SESSION CHAIRED BY BILL BREWER SENATE HOUSE VOLUME CXXI EDITED BY GUY LONGWORTH UNIVERSITY OF LONDON proceedings of the aristotelian society 141st session issue no. 2 volume cxxi 2020–2021 the linguistic approach to ontology lee walters university of southampton monday, 18 january 2021 17.30–19.15 online via zoom Please visit our website for further details contact [email protected] www.aristoteliansociety.org.uk © 2021 the aristotelian society biography Lee Walters is an Associate Professor in Philosophy at the University of Southampton. Prior to joining Southampton, Lee studied philosophy at UCL and taught at Oxford. Lee’s main interests are in metaphysics, the philosophy of language, and philosophical Logic, with a particular emphasis on the philosophy of fiction. Lee has been an Associate Editor of Analysis; a trustee of the British Society of Aesthetics; has held a British Academy Mid-Career Fellowship; and has been a junior fellow at the Institute for Advanced Study, CEU, Budapest. editorial note The following paper is a draft version that can only be cited or quoted with the author’s permission. The final paper will be published in Proceedings of the Aristotelian Society, Issue No. 2, Volume CXXI (2021). Please visit the Society’s website for subscription information: aristoteliansociety.org.uk. The Linguistic Approach to Ontology 1. Introduction Following Quine (1948), and more recently Hofweber (2016), I take the central ontological question to be what is there? in a sense to be precisified in §3. This contrasts with some recent approaches to ontology that take the central question to be what is fundamental? It is, of course, interesting to explore relations of dependence and supervenience between different sorts of entity, but this endeavour of explaining what depends on what, is an investigation of relations within our ontology. And in order to fully explore these relations we need to determine what our ontology consists of, namely what there is. In any case, there is room for both projects.1 What I want to do here is explore the prospects for the Linguistic Approach to (Meta)Ontology: we determine what there is by examining how natural languages work. In §§2-3, I outline two linguistic approaches to ontology due to Frege and Hofweber, respectively. I reject such approaches (§4) for the familiar reason that empty names can figure in true sentences. Still there is a kernel of truth in both approaches. The truth that these approaches encapsulate is often thought to be the claim that true subject-predicate sentences are ontologically committing. This thought is appealing, but it is ultimately untenable. There is, however, a truth in the area namely that true property ascriptions are ontologically committing. In §§5-6 I pull these two claims apart, defending the latter and rejecting the former. 1 Why not take the ontological question to be what exists? Well on the intended readings, I think this is equivalent to the question of what is there? The English word exist suffers from the drawback that we talk of events occurring or happening rather than existing (Williamson), but events are part of ontology – there are events, after all. We also say that Socrates no longer exists, but eternalists will want to include him in our ontology. Here, however, exists seems to be on all fours with there is since we do not say that Socrates is, but only that he used to be. 1 The package of these views faces two related challenges: what are predicates doing in subject- predicate sentences, and what accounts for the semantic profile of empty names in the absence of a referent. In §6 I make a start on answering this. §7 briefly mirrors the previous debates concerning first-order ontology in the second-order case. The picture we are left with is one on which we need to determine which sentences ascribe properties to objects, and there does not appear to be any formal criterion for this. Before commencing with the substance of the paper, let me say something about what I am doing here and the paper’s ambitions. This paper is wide-ranging and rests on controversial claims at many points, claims that cannot in the space available be defended in detail (I may have lost some readers already, simply by claiming that there are true sentences containing empty names). I make no apology for this. The issues that I want to address are wide-ranging and touch on many areas of philosophy, and no one paper could deal with all of them adequately. There are arguments in this paper, but I think reader will find it most fruitful if they take me to be sketching what I take the best approach to these issues to be, one that is intuitive, but nevertheless seeks to address the hard questions that face such a view. Even if you ultimately reject the view advocated here, I think there is still something to be learned by engaging with it and the debate it seeks to illuminate. 2. Fregeanism A particularly influential approach to ontology is that inspired by Frege (1953). Fregeanism can be characterised for our purposes as the conjunction of the following two claims (cf. MacBride, 2003: 108). 2 Syntactic Decisiveness: if an expression exhibits the characteristic syntactic features of a singular term, then that fact decisively determines that the expression in question has the semantic function of a singular term, i.e. the function of referring to a single object. Following Hofweber (2016), let us call expressions that exhibit the syntactic features of a singular term, syntactic singular terms, and expressions that have the function of referring to a single object, whether or not they carry out that function, semantically singular terms. Syntactic Decisiveness claims that all syntactically singular terms are semantically singular terms. Referential Minimalism: the mere fact that a semantically singular term features in a true sentence (of a certain sort) determines that this expression refers to a single object.2 Together these two claims entail that true sentences (of a certain sort)3 containing syntactically singular terms are ontologically committing, since they commit us to the objects referred to by the syntactically singular terms in these sentences. If we conceive of existential quantification into (syntactically) singular term position (nominal quantification) as quantification over a domain of objects, then true sentences that contain such quantification will also be ontologically committing, since such sentences can only be true if there is an object that is a witness of the quantificational claim. 2 I am using reference and its cognates, as is standard in analytic philosophy, to pick out a relation between a singular term and an object. We should distinguish reference from aboutness. On the assumption that ‘Pegasus’ is an empty name, ‘Pegasus’ does not refer to anything and I cannot refer to Pegasus. Nevertheless, there are myths about Pegasus and I can think about Pegasus, or so it very much seems. Here I follow Crane (2013: 8-10). 3 I shall leave this qualification implicit until I return to it in §5. 3 So, a central ontological question for the Fregean is to determine the true sentences containing syntactically singular terms or an existential nominal quantifier. Once we have done this, our (first-order) ontological commitments follow. 3. Hofweber’s Rejection of Fregeanism Hofweber (2016) has recently outlined an interesting alternative linguistic approach to the Fregeanism sketched above. Hofweber argues that there are two sorts of syntactically singular terms, since some but not all syntactically singular terms are semantically singular terms. That is, some but not all syntactically singular terms have the function of referring. Hofweber thus departs from the Fregean by rejecting Syntactic Decisiveness. For example, Hofweber argues that there are certain puzzling features of our number talk that are best explained by taking neither ‘the number of moons of Jupiter’ nor ‘four’ as used in (1) to be have the function of referring. 1. The number of moons of Jupiter is four. Furthermore, just as Hofweber distinguishes between two sorts of syntactically singular term, he also distinguishes between two readings of nominal quantifiers External quantification: nominal quantifiers range over a domain of objects as assumed in §2. Internal quantification: nominal quantifiers generalize into syntactic position and so have an inferential role or substitutional reading. 4 External quantification is governed by a free logic. That is, we can only move from Fa to ∃x Fx on the assumption that ‘a’ refers (similarly from ∀x Fx to Fa). This is because, some syntactically singular terms are not semantically singular, they do not refer. Internal quantifiers, on the other hand, behave classically. That is, we can conclude that there is something (internally) that is F (Σx Fx) from Fa and similarly from the claim that everything internally is F (Π Fx) to Fa. If Hofweber is right about number words, then generalizations about numbers – (there are) some numbers (that) are prime/perfect/larger than 10 – cannot be accounted for in terms of external quantification because none of the objects that the external quantifiers range over are identical to four, for example, because four is not a referring term. And a result, on Hofweber’s picture we need an internal reading of the quantifier. But couldn’t we make do with only internal quantifiers? No, unreferred to objects mean we need an external nominal quantifier, since there are no true sentences from which to reach 2. There are some objects that will never be referred to which would be required if ‘some objects’ in (2) were an internally quantified noun phrase - Σx Fx is true iff there is some (externally) substitution instance of x ‘a’ such that Fa.